Open Research OnlineThe Open University’s repository of research publicationsand other research outputs

Performance comparison of inkjet and thermal transferprinted passive ultra-high-frequency radio-frequencyidentification tagsJournal ItemHow to cite:

Kgwadi, Monageng; Rizwan, Muhammad; Adhur Kutty, Ajith; Virkki, Johanna; Ukkonen, Leena and Drysdale,Timothy D. (2016). Performance comparison of inkjet and thermal transfer printed passive ultra-high-frequencyradio-frequency identification tags. IET Microwaves, Antennas & Propagation, 10(14) pp. 1507–1514.

For guidance on citations see FAQs.

c© 2016 The Institution of Engineering and Technology

https://creativecommons.org/licenses/by-nc-nd/4.0/

Version: Accepted Manuscript

Link(s) to article on publisher’s website:http://dx.doi.org/doi:10.1049/iet-map.2016.0331

Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyrightowners. For more information on Open Research Online’s data policy on reuse of materials please consult the policiespage.

oro.open.ac.uk

Performance Comparison of Inkjet and Thermal Transfer Printed

Passive UHF RFID Tags

Monageng Kgwadi1, Muhammad Rizwan2, Ajith Adhur Kutty2, Johanna Virkki2, Leena Ukkonen2,

Timothy D. Drysdale3,*

1Electronics and Nanoscale Engineering Division, University of Glasgow, G12 8QQ, UK.2Department of Electronics and Communications Engineering, Tampere University of Technology,

Tampere, Finland.3Department of Engineering and Innovation, The Open University, Milton Keynes, MK7 6AA,

UK.*tim.drysdale@open.ac.uk

Abstract: We compare the maximum read range of passive ultra high frequency (UHF) radio fre-

quency identification (RFID) tags that have been produced using different metal printing tech-

niques, specifically inkjet printing and thermal transfer printing. We used the same substrate

(THERMLfilm), antenna designs, and electronic circuitry in our comparison so as to isolate the

effect of the metal printing. Due to the high metal conductivity, the thermal transfer printed tags

printed with copper ink performed as well or better than the inkjet printed tags printed with silver

ink, even when we changed the inkjet printed tags to a Kapton substrate that is better suited to

inkjet printing. The aluminium thermal transfer printed tags had up to 33% less read range than

copper thermal transfer printed tags. Thermal transfer printing needs no sintering, and provides an

attractive alternative low-cost fabrication method. Characterisation of the printed traces by both

methods reveals that the printing techniques achieve similar surface roughness between 19.8 nm

and 21.2 nm RMS. The achieved conductivities for thermal transfer printing on THERMLfilm

were 2.6⇥107 S/m and 3.9⇥107 S/m for aluminium and copper films respectively while inkjet

printing achieved 1.7⇥107 S/m conductivity on the same substrate. The best measured read range

1

for THERMLfilm was 10.6m . Across the different tag designs, the measured read ranges were 15-

60% (1-10%) better for thermal printing, compared to inkjet printing on THERMLfilm (Kapton).

1. Introduction

The drive to develop low-cost, flexible and lightweight electronics to support Internet of Things

(IoT) applications has received a lot of attention in recent years [1]. The IoT is envisaged to

consist ultimately of trillions of communicating devices operating in diverse applications such as

smart cities, personal area networks, smart homes, supply chain management, and smart grids

[2, 3, 4, 5] to name but a few. Pervasive adoption of IoT application hinges on a low unit cost for

the nodes [6], hence the need to develop low-cost manufacturing of electronics by both academia

and industry [3, 4].

A number of inexpensive techniques are available for printing electronics. These include addi-

tive methods such as inkjet printing (e.g. of antennas [7], frequency selective surfaces [8], UHF

passive components[9], sensors [10] and circuits [11]) and screen printing (e.g. of electronic cir-

cuits [12] and passive sensors [13]). Both are well known and widely used approaches to deposit

an ink containing conductive particles or flakes typically metallic, which then requires sintering

to achieve the maximum available conductivity. Nonetheless, newer developments may from time

to time emerge, potentially offering improved processing methods that are faster, more convenient

and less expensive. One such development is the introduction of metal ribbons for thermal transfer

printing (TTP), which has enabled the printing of antennas [14] and frequency selective surfaces

[15] without the need for a sintering step (which is essential in inkjet printing). On the other hand,

the metal thickness available through thermal transfer printing appears relatively thin (ca. 300 nm)

compared with inkjet printed metal thicknesses which can be over 1µm. This difference would

appear to be critical to the operation of a UHF RFID tag where the skin depth is ⇠2µm in pure

copper at 1 GHz [16], yet the quality of the deposited metal (represented by its conductivity or

sheet resistance) is also a determining factor, as is the influence of substrate on the adhesion of the

ink, surface roughness, loss and mechanical flexibility [17, 18]. In this paper, we seek to compare

2

the overall performance obtained by designing a set of three UHF RFID tags with different anten-

nas that use a representative mixture of thick and thin line widths, printing them using both inkjet

and thermal transfer printing on the same substrate, and then determining from measurements, the

maximum read range. In this paper we confine our interest to a subset of printing techniques (inkjet

and thermal transfer) that can support individually customised designs, so we do not consider the

relative performance of screen printing. Inkjet and screen printing have already been benchmarked

against each other, e.g. [19]. To the best of our knowledge, this is the first systematic comparison

of thermal transfer printed UHF RFID antennas to either.

The paper is organised as follows; the printing techniques and the equipment used in this study

are discussed in Section 2. Physical and electrical characterisation of the traces of the two methods

are discussed in Section 3. The UHF RFID tag designs used in this study and the evaluation method

are discussed in Section 4. Section 5 presents the simulated read range of the tags while Section 6

presents results of the measurements. Conclusions are then drawn in Section 7.

2. Methods

We wish to compare the effect of changing the technique for printing the metal antennas and cir-

cuit traces, so we keep the rest of the tag as similar as possible. In particular, to use the same

antenna designs, the same electronic circuit, and the same substrate. The use of the same substrate

allows us to eliminate sources of variation arising from differences in the thickness, dielectric con-

stant, dielectric loss, and surface roughness. The only substrate available to us that could achieve

consistent results using both techniques was THERMLfilm SELECT 21944E, formerly known as

TC-390, from Flexcon.

For brevity, we refer to this substrate as THERMLfilm in the rest of the paper. THERMLfilm

is a 50µm thick polyester based flexible substrate with a glossy finish backed with a 20µm thick

acrylic adhesive and a removable 56µm thick glassine liner suitable for TTP [20]. The relative

dielectric constant of THERMLfilm SELECT 21944E is ✏r

= 3.2 [15]. THERMLfilm is best

3

suited to thermal transfer printing, and has a porous surface that absorbs a fraction of the ink when

deposited using inkjet printing. This leads to the spreading of the ink drops and thinning of the

metalisation which is expected to slightly degrade the performance of the inkjet printed tags, as

discussed in Ref. [21]. So we can estimate the impact of the THERMLfilm porosity on inkjet

printing, we also inkjet print onto Kapton, which has negligible absorption and is well suited to

inkjet printing.

2.1. Inkjet Printing

For this work, we used a Fujifilm Dimatix DMP-2831 material inkjet printer, with 10 pL print

nozzles. The ink was Harima’s NPS-JL silver nanoparticle ink with particle sizes of 5-12 nm

with a maximum achievable conductivity in the range of 1.6⇥107 - 2.5⇥107 S/m [22]. We chose

this ink because it is commonly used, but we note other (potentially more cost-effective) metallic

inks are available and under development [23]. The optimized inkjet printing key parameters are

presented in Table 1. Four of the sixteen jets were used simultaneously to obtain continuous traces

on THERMLfilm at a jetting voltage of 28 V and a cartridge temperature of 40 oC. A high platen

temperature (60 oC) was used to reduce ink drying time. After printing, the sample was left on the

platen for between two and three minutes so as to partially dry the ink. The samples were then

sintered for one hour at 150 oC.Table 1 Inkjet Printing Key Parameters

Parameter valuePlaten Temperature 60 oC

Cartridge Temperature 40 oCMax. Jetting Frequency 9 KHz

Jetting Voltage 28 VDrop Velocity 8-9 m/sDrop Volume 10 pL

Drop Spacing/Pattern resolution 40 µm/635 DPI

2.2. Thermal Transfer Printing

Thermal transfer printing employs a multi-layered ribbon containing a prepared ink layer. For

metal printing, the ribbon features a plastic membrane on top of which a thin layer of metal is

4

bonded using a resin, and a heat-sensitive acrylic adhesive on top of the metal. Printing is achieved

by using a thermal print-head to selectively activate the desired regions of the heat-sensitive adhe-

sive which then adheres to a substrate placed in physical contact with the ribbon and in the process,

transfers the thin metal from the ribbon to the substrate as shown by Fig. 1a. We use two different

IIMAK MetallographTM Conductive Thermal Transfer Ribbons (CTTR). One ribbon had a 260 nm

thick aluminium film while the other has 340 nm thick copper film and both have a 1 µm thick

heat-sensitive acrylic adhesive. A schematic of the Zebra S4M thermal printer used is shown in

Fig. 1b showing the operation of the printer, while a photograph of the printer is shown in Fig. 1c.

The printer has a print-head resolution of 300 dpi and a print speed of 5 cm/s. The sample can be

used immediately with no drying or curing required.

Table 2 Metallograph ribbon : key parameters

Metal Aluminium CopperMetalisation thickness 260 nm 340 nm

Adhesive thickness 1µm 1µm

3. Print Characterisation

We note that multi-layer printing techniques are available for each type of printing, but are not

equally mature. Hence we confine our present study to single layer printing. Before producing the

UHF RFID tags, we characterised traces printed using both methods.

3.1. Inkjet Printing

Inkjet printing produces continous traces on THERMLfilm when using the parameters in Table

1. As shown in Fig. 2a , a white light microscope image of the printed trace shows some small

discontinuities in the printed traces, which could be caused by air bubbles forming during jetting,

dust or sintering artifacts. Current scattering due to surface roughness is known to cause up to

40% increase the resistivity of thin films [24], which inevitably results in increased ohmic losses

in antennas, and lower read ranges for the UHF RFID tags that we consider in this paper. A

surface roughness measurement of a trace was made using a Bruker Dimension Icon atomic force

5

a

b c

Fig. 1. Thermal printing processa Schematic of the printing methodb Schematic of the Zebra S4M printerc Photograph of the Zebra S4M printer

microscope (AFM) on an area of 5⇥2 µm that was representative of the sample and is shown

in Fig. 2b. The measured root mean square (RMS) surface roughness was 21.2 nm, which is

comparable to the measurements in [19] of 22.1 nm (RMS) inkjet printing on Polyetherimide (PEI),

a substrate well suited for inkjet printing. THERMLfilm substrate has a measured surface rough-

ness of 26.5 nm (RMS).

The thickness of the metal traces was measured using a Veeco Wyko NT1100 optical profiler

to be in the range of 0.40-0.60µm. The measured sheet resistance of the silver inkjet traces on

6

a b

Fig. 2. Physical characteristics of thermal transfer printed tracesa Microscope image of inkjet trace on THERMLfilmb AFM image of inkjet trace on THERMLfilm. Note that the trace edges are outside the edge ofthe pictures, and that the AFM trace covers ⇠ 1/100th of the the area shown

THERMLfilm using the four probe method on Agilent’s B1500 Semiconductor Device Analyzer

with American Probe & Technologies’ Quasi Kelvin Probes was 0.15 ⌦/⇤. We calculated con-

ductivity using the average metal thickness and sheet resistance to be 1.7⇥107 S/m. The calculated

skin depth for inkjet on THERMLfilm at 900 MHz is 4.1µm. For comparison, the realized metal

thickness on Kapton was 1.0µm while the measured sheet resistance was 0.056 ⌦/⇤, yielding

conductivity of 1.8⇥107 S/m. The skin depth of inkjet on Kapton is 3.9µm at 900 MHz. Thus

the metalisation is 12.9% of the skin depth for inkjet printed traces on THERMLfilm, while it

is 25.6% on Kapton. Therefore larger ohmic losses are expected on the inkjet printed traces on

THERMLfilm than on Kapton.

3.2. Thermal Transfer Printing

White light optical microscope images of thermal transfer printed metal traces are shown in Fig.

3 with aluminium in Fig. 3a and copper in Fig. 3b. The images photos show hairline disconti-

nuities on both the aluminium and copper traces. The discontinuities are more clearly shown in

the aluminium trace due to the contrast in the picture but the density is higher in the copper traces

(⇠ 8/mm2). We attribute these discontinuities to mechanical stresses associated with the handling

of this prototype material, although further investigations are required to characterise the mecha-

nism and are outside the scope of this paper.

7

The sheet resistance of the thermally printed traces on THERMLfilm was measured by the four

probe method. The results are summarised in Table 3. The sheet resistance measured for both

copper and aluminium traces shows a small dependence on the direction of print. This is attributed

to the print elements being positioned in a single line across the width of the print substrate, per-

pendicular to the direction of substrate movement. The aluminium traces have a measured sheet

resistance of 0.16 ⌦/⇤ along print direction and 0.15 ⌦/⇤ across the print direction. Copper traces

have a measured sheet resistance of 0.10 ⌦/⇤ along the print direction and 0.08 ⌦/⇤ across the

print direction. Thus the highest conductivity of the aluminium and copper traces are 2.6⇥107 S/m

and 3.9⇥107 S/m respectively. The calculated skin depth at 900 MHz for aluminium and copper

on THERMLfilm is 3.4µm and 3.1µm respectively, yielding a metalisation thickness that is 8.4 %

and 12.6 % of the skin depth respectively. Therefore the copper tags are expected to have a higher

read range than the aluminium tags.

An AFM surface roughness measurement of both traces was made on an area of 5⇥5 µm.

The surface roughness of the aluminium and copper traces are shown in Fig. 4a and Fig. 4b

respectively. The aluminium traces had an RMS roughness of 20.8 nm while the copper traces

have an RMS roughness of 19.8 nm. The copper traces have a marginally smoother surface than

aluminium with 0.4 nm average (1.0 nm rms). Both the aluminium and copper traces have a

slightly smoother surface roughness than the inkjet traces on THERMLfilm (21.2 nm) and the

surface roughness reported in [19].

Table 3 Key physical and DC characteristics of inkjet and TTP. Note that TTP sheet resistance and conductivityresults are reported both perpendicular (?) and parallel (k) to the direction of substrate movement

Inkjet TTP Aluminium TTP Copper InkjetTHERMLfilm THERMLfilm THERMLfilm Kapton

Metal Thickness 400-600 nm 260 nm 340 nm 1µmSurface Roughness 21.2 nm 20.8 nm 19.8 nm -

(RMS)Sheet Resistance 0.15 ⌦/⇤ 0.15 ⌦/⇤ ? 0.08 ⌦/⇤ ? 0.056 ⌦/⇤

- 0.16 ⌦/⇤ k 0.10 ⌦/⇤ k -Conductivity 1.7⇥107 S/m 2.6⇥107 S/m 3.9⇥107 S/m 1.8⇥107 S/mSkin Depth 4.1µm 3.4µm 3.1µm 3.9µmat 900 MHz

8

a b

Fig. 3. Microscope images of printed traces on THERMLfilm substrate.a Aluminium traceb Copper trace

a b

Fig. 4. AFM images showing the surface roughness of thermal-transfer printed traces onTHERMLfilm.a Aluminium trace surface morphologyb Copper trace surface morphology

4. RFID tag design

Three different UHF RFID tag antenna designs were chosen to highlight the benefits and limita-

tions of the printing techniques. As shown in Fig. 5, the antenna designs have different geometries

featuring wide and narrow tracks that experience different conductor losses and are expected to

have different read ranges. By using three different tags, we reduce the influence of any individual

antenna design upon the results. In all cases, we used only a single printing pass.

NXP UCODE G2iL series RFID integrated circuits (IC) were attached to each of the fabricated

antennas using conductive silver epoxy (Circuit Works CW2400) on the manufacturer-mounted

9

a

b

c

Fig. 5. Schematics of the RFID tags. All dimensions in millimetres.a Tag design 1b Tag design 2c Tag design 3

fixture with two 3x3 mm2 pads. After the attachment, the tags were placed in oven from 10 min-

utes at 70 �C to dry the silver epoxy. The read range of the tags was measured in a 120x80x80 cm

Voyantic RFID measurement cabinet (Anechoic Chamber) using a Voyantic Tagformance RFID

measurement unit. The equipment calculates the theoretical read range d from the backscattered

signal from the tag assuming a reflection-less environment :

d =�

4⇡

rEIRP

Pth

Liso

, (1)

where � is the wavelength of the transmitted signal to the tag, Pth

is the measured threshold power

10

required to activate the tag, Liso

is the measured path loss, and EIRP is the effective isotropically

radiated power which is limited to 3.28 W according to the European regulations [25]. The RFID

antenna reader has a linearly polarised patch antenna with 8 dB gain operating at 800-1000 MHz.

The reader’s output power ranges from 0 dBm to 27 dBm and the RF receiver has sensitivitiy of

-75 dBm. The power was ramped up in 0.1 dBm steps until the tag is activated.

5. Simulation Results

The RFID antennas mentioned in the previous section were characterised using ANSYS high fre-

quency structure simulator (HFSS) to estimate their read range. The material properties outlined

in Table 3 were used to simulate the performance of each antenna design. THERMLfilm and Kap-

ton films were modelled as 50µm thick substrates with relative dielectric constants of 3.2 and 3.0

respectively. The conductive tracks of the antennas were modeled using the layered impedance

boundary in HFSS. A lumped port with 50⌦ source impedance was used as excitation at antenna

terminals. Read range of the tag was computed using the expression,

Rrange

=⇣ �

4⇡

⌘s⌧ ·GTg

· EIRP

Pchip

, (2)

where � is the wavelength, ⌧ is the power transmission coefficient, GTg

is the gain of the tag,

EIRP is the equivalent isotropic radiated power and Pchip

is the minimum power required to wake

up the microchip. The power transmission coefficient depends on the antenna input impedance and

the input impedance of the microchip as shown by the expression,

⌧ =4R

A

RC

(RA

+RC

)2 + (XA

+XC

)2, (3)

where RA

+ jXA

is the complex input impedance of the antenna and RC

+ jXC

is the complex

input impedance of the chip. The equivalent circuit model of the NXP microchip developed in [27]

was used to obtain the chip impedance across various frequencies.

11

The read range estimates computed using Equation (2) are plotted against frequency as shown in

Fig. 6a, 6b, and 6c for tag design 1, design 2 and design 3 respectively. From the simulations, inkjet

printed silver and thermal transfer printed aluminium on THERMLfilm substrate are expected to

have very similar read range for all the tag designs. Inkjet printed silver on Kapton surface has the

highest estimated read range, which is due to its higher thickness though its conductivity is lower.

Thermal printed copper is expected to have a comparable albeit lower read range to inkjet printed

silver on Kapton.

6. Measurement Results

The average theoretical read range of tag design one fabricated by the two printing methods is

presented in Fig. 7. TTP copper antennas on THERMLfilm achieve a maximum read range of

6.6 m, while TTP aluminium antennas achieve 4.4 m read range. The difference of 2.2 m in the

read range between the copper and aluminium TTP antennas is attributed to the higher conductivity

of the copper traces and a thicker metalisation profile. These properties together with the narrow

tracks of the antenna design results in significant conductor losses in the aluminium antennas and

results in shorter average read range. The inkjet printed antennas on Kapton achieves a maximum

read range of 5.9 m at 930 MHz. Inkjet printed antennas of THERMLfilm achieves a read range of

4.2 m, which is a 1.7 m less than the baseline performance. The difference is due to the suboptimal

printing surface which results in thinner metalisation profile due to absorption of ink. The narrow

metal tracks of the antenna design magnify the effects of the thin metalisation by increasing the

ohmic losses, resulting in the significant reduction in reading range. The copper TTP antennas

manages an improvement of 0.8 m in read range over the inkjet antenna on Kapton.

The average read ranges of antenna design two are shown in Fig. 8. The printed examples all

achieve relatively similar read range, which we attribute to the thick lines of the design reducing

the conductor losses and masking the difference in conductivity. Inkjet printed tags on THERML-

film achieves 7.9 m read range which is 0.7 m less than the baseline due to the afore-mentioned

suboptimal substrate properties. Aluminum TTP antennas on THERMLfilm achieve a read range

of 7.7 m, which is similar to inkjet on THERMLfilm with a difference of 0.2 m. The copper TTP

12

a

b

c

Fig. 6. Simulated read ranges for antennas using material properties in Table 3a Tag design 1b Tag design 2c Tag design 3

13

Fig. 7. Measured theoretical reading range of tag design 1 fabricated by inkjet and thermal trans-fer printing methods

antennas on THERMLfilm achieve peak read range of 9.0 m which is 0.4 m more than the inkjet

printed antenna on Kapton. The inkjet on Kapton antennas achieve a maximum read range of 8.6 m

at 960 MHz. The performance improvement of TTP copper over the baseline inkjet read range is

due to the higher conductivity of the copper traces (see Table 3).

The average read range of tags of design three are shown in Fig. 9. There is more spread in

the read range (2.3 m), similar to design one (1.2 m) due to the narrow and longer metal traces

highlighting the effects of the conductor losses. Inkjet printed tags on THERMLfilm achieve a

maximum read range of 9.2 m which is 1.6 m less than the baseline due to the suboptimal sub-

strate. TTP aluminium antennas on THERMLfilm achieve a peak read range of 9.8 m, while the

TTP copper antennas on THERMLfilm achieve a peak read range of 10.6 m. The inkjet tags on

Kapton achieves a maximum read range equal to TTP copper tags on THERMLfilm of 10.6 m at

930 MHz.

A summary of the measured maximum read ranges is presented in Table 4. The achieved maxi-

mum theoretical read ranges by aluminium TTP antennas is equal to inkjet printing on THERML-

film for the three designs in this study. The copper ribbon achieves better read range due a higher

conductivity and a thicker metalisation compared to the aluminium ribbon leading to lower con-

14

Fig. 8. Measured theoretical reading range of tag design 2 fabricated by inkjet and thermal trans-fer printing methods

Fig. 9. Measured theoretical reading range of tag design 3 fabricated by inkjet and thermal trans-fer printing methods

ductor losses, which gives an overall performance similar to inkjet on Kapton. The simulation

results suggest that inkjet printing on Kapton could be slightly better than thermal transfer print-

ing with copper, but these simulations do not take into account all of the fabrication parameters

due to modelling the metallised regions with idealised smooth surfaces and edges. In the mea-

surements, where all practical effects are included in the results, the thermal transfer printed tags

outperform the inkjet printed labels by between 15-60 % (1-10 %) compared to inkjet printing on

THERMLfilm (Kapton).

15

Table 4 Measured Theoretical Read Range

Printing Method Measured Theoretical Read range (m)tag design 1 tag design 2 tag design 3

and substrate (940 MHz) (960 MHz) (920 MHz)Thermal TransferPrinting (Al)

THEMLfilm 4.4 7.7 9.8Thermal TransferPrinting (Cu)

THEMLfilm 6.6 9.0 10.6Inkjet (Ag)

THEMLfilm 4.2 7.9 9.2Kapton 5.8 8.6 10.6

7. Conclusion

We fabricated UHF RFID tags and determined the maximum read range from measurements. By

using different metal printing techniques, but keeping the tag antenna designs, electronic circuitry

and substrate the same, we were able to separate out the influence of the printing method on the

overall performance. We were primarily interested to see whether thermal transfer metal printing

could produce acceptable performance, compared with inkjet printing. Our results indicate that

thermal transfer printing using copper is able to equal or outperform inkjet printing with silver

inks, regardless of whether the ink jet printing was conducted on the same substrate as the thermal

printing, or on a substrate specially developed for inkjet printing. Our study used three differ-

ent tag designs, representing antennas with a mix of thick and thin features, and more than one

example was printed in most cases. Across the three designs, the thermal transfer printed labels

outperform the inkjet printed labels by between 15-60 % (1-10 %) compared to inkjet printing on

THERMLfilm (Kapton). Therefore our results justify further investigation of the use of thermal

transfer printing in the production of UHF RFID tags in specific, and also indicate that it potentially

has a role to play in printed electronics more generally. Thermal transfer printing is an attractive

alternative to inkjet printing because it does not require sintering, thus eliminating an additional

step from the tag fabrication process, saving time and money. There are opportunities for future

work in the area of multiple-pass thermal transfer printing, establishing the fine resolution printing

limits, and automating the assembly of printed structures that incorporate electronic elements, such

16

as switching elements that permit antenna reconfigurability [26].

Acknowledgment

The authors would like to thank Daniel Harrison from IIMAK for donation of both the aluminum

and copper MetallographTM ribbons. This work was partly supported by European Commis-

sion from the Seventh Framework Programme through (FP7) AdvIOT (www.adviot.eu) and the

Academy of Finland, Emil Aaltonen Foundation and TEKES.

8. References

[1] Tan, L. and Wang, N.: ‘Future Internet: The Internet of Things’, Adv. Comp. Theory and Eng.

(ICACTE), Aug. 2010 3rd Int. Conf., Chengdu, China, Aug. 2010, 5, pp. 476-380.

[2] Harmon, R.R., Castro-Leon, E.G., Bhide, S.: ‘Smart cities and the Internet of Things,’ Man.

of Eng. and Tech. (PICMET), Portland, OR, USA, Aug 2015, pp. 485-494,

[3] Miorandi, D., Sicari, S.,De Pellegrini, F., Chlamtac, I.: ‘Internet of things: Vision, applications

and research challenges’, Ad Hoc Networks, 2012, 10,(7), pp. 1497-1516.

[4] Stankovic, J.A., ‘Research Directions for the Internet of Things,’ Internet of Things Journal,

IEEE , 2014, 1, (1), pp. 3-9,

[5] Gubbi, J., Buyya, R., Slaven Marusic, S., Marimuthu Palaniswami, M.: ‘Internet of Things

(IoT): A vision, architectural elements, and future directions’, Future Gen. Comp. Sys., 2013,

29, (7), pp. 1645-1660,

[6] Wu, G., Talwar, S., Johnsson, K., Himayat, N., Johnson, K.D.: ‘M2M: From mobile to embed-

ded internet,’ Comm. Mag., IEEE , 2011, 49, (4), pp.36-43.

[7] Mantysalo, M., and Mansikkamaki, P.: ‘An Inkjet-Deposited Antenna for 2.4GHz Applica-

tions’, AEU- Int Jrnl. of Elec. and Comms., 2009,63,(1), pp. 31-35.

17

[8] Batchelor, J.C., Parker, E.A., Miller, J.A., Sanchez-Romaguera, V., Yeates, S.G.: ‘Inkjet print-

ing of frequency selective surfaces,’ Electronics Letters,2009, 45, (1), pp. 7-8

[9] Redinger, D., Molesa, S., Shong Y., Farschi, R., and Vivek, S.: ‘An ink-jet-deposited passive

component process for RFID,’ Electron Dev., IEEE Trans., 2004 , 51, (12), pp. 1978-1983.

[10] Kim, S., Cook, B., Taoran L., Cooper, J., Hoseon L., Lakafosis, V., Vyas, R., Moro, R.,

Bozzi, M., Georgiadis, A., Collado, A., Tentzeris, M.M.: ‘Inkjet-printed antennas, sensors and

circuits on paper substrate,’ Microwaves, Ant. & Pro., IET, 2013 , 7, (10), pp. 858-868.

[11] Virtanen, J., Virkki, J., Ukkonen, L., Sydanheimo, L.: ‘Inkjet-Printed UHF RFID Tags on

Renewable Materials,’Adv. Internet of Things, 2012, 2 (4), pp. 79-85.

[12] Zhou, J., Ge T., Ng, E., Chang, J.S.: ‘Fully Additive Low-Cost Printed Electronics With Very

Low Process Variations,’ Electron Devices, IEEE Trans., 2016, 63, (2), pp. 793-799.

[13] Merilampi, S., Bjorninen, T., Ukkonen, L., Ruuskanen, P., Sydanheimo, L.,:’Embedded wire-

less strain sensors based on printed RFID tag’, Sensor Review, 2011, 31,(1) , pp. 32-40.

[14] Kgwadi, M., Vourch, C.J., Harrison, D.J., Drysdale, T.D.: ‘On-demand printing of antennas

for TV white-space communications,’ Ant. and Prop. Conf. (LAPC), Loughborough, UK., Nov.

2014, pp. 553-556.

[15] Whyte, G., Harrison, D., Cumming, D., and Drysdale, T.: ‘Direct Printing of Flexible Metal-

lic Millimetre-wave Frequency Selective Surfaces’, Anten. and Prop. Soc. Int. Symp. (AP-

SURSI), Toronto, ON, Canada, July 2010, pp. 1-4.

[16] Arora, N. D. :‘Modeling and characterization of copper interconnects for SoC design,’ Sim.

of Semiconductor Processes and Devs., SISPAD 2003. Intl. Conf., Boston, MA, USA, 2003,

pp. 1-6.

[17] Merilampi, S. L., Bjorninen, T., Ukkonen, L., Ruuskanen, P., and Sydanheimo,

L.:‘Characterization of UHF RFID tags fabricated directly on convex surfaces by pad print-

ing’ , Int. Jrnl. of Adv. Manuf Technol, 2011, 53(5), pp. 557-591.

18

[18] Hicks, W. T., Allington, T. R., and Johnson,V.: ‘Membrane Touch switches: Thick-Film

Materials Systems and Processing Options,’ IEEE Trans. on Components, Hyb., and Manuf.

Tech.,1980, 3,(4), pp. 518-524,

[19] Salmeron, J.F., Molina-Lopez, F., Briand, D., Ruan, J.J., Rivandeneyra, A., Carvajal, M.A.,

Capitan-Vallvey, L.F., Derooij, N.F., and Palma, A.J.:’Properties and Printability of Inkjet and

Screen-Printed Silver Patterns for RFID Antennas’ , Jrnl. of Elect. Materials, 2013, 43(2), pp.

604-617 .

[20] ‘Product Construction Sheet’, http://www.uce-druckservice.de/downloads/flexcon/select/THE

RMLfilm%20Select%2021944E.pdf, accessed 20 January 2016

[21] Virkki, J., Virtanen, J., Sydanheimo, L., Ukkonen, L., Tentzeris, M.: ‘Embedding inkjet-

printed antennas into plywood structures for identification and sensing,’ RFID-Tech. and App.

(RFID-TA)IEEE Intl. Conf. , Nice, France, Nov. 2012, pp. 34-39.

[22] ‘Nano Paste Series’, https://www.harima.co.jp/en/products/electronics/pdf/brochure16e 23

.pdf, accessed 20 January 2016

[23] Ren, Y., Virkki, J., Sydanheimo, L., Ukkonen, L., ‘Optimisation of manufacturing parameters

for inkjet-printed and photonically sintered metallic nanoparticle UHF RFID tags,’ Elec. Let.,

2014, 50, (21), pp. 1504-1505.

[24] Timoshevskii, V., Ke, Y., Guo, H., Gall, D.:’The influence of surface roughness on electrical

conductance of thin Cu films: An ab initio study’ , Jrnl. of Applied Phys., 2008 103.

[25] ‘ETSI TR 101 445 v1.1.1 (2002-04)’ http://www.etsi.org/deliver/etsi tr/101400 101499/1014

45/01.01.01 60/tr 101445v010101p.pdf, accessed 14 February 2016.

[26] Kgwadi, M., Drysdale, T. D.: ‘Diode-switched thermal-transfer printed antenna on flexible

substrate’, IET Electronics Letters, 2016, 54, (4), pp. 258-260.

[27] Bjorninen, T., Syndanheimo, L., Ukkonen, L.: ‘Development and validation of an equivalent

circuit model for UHF RFID IC based on wireless tag measurements’, Digest of AMTA Symp.,

Bellevue, WA, USA, Oct. 2012, pp 480-485.

19